Reprint requests should be addressed to Stephen Nowicki, Jr., Department of Psychology,. Emory University ..... Encounters. Indianapolis, IN: Bobbs-Merrill.
JOURNAL
OF RESEARCH
IN PERSONALITY
25,
322-333 (1991)
Interpersonal Complementarity and Time of Interaction in Female Relationships STEPHEN NOWICKI,
JR., AND SUSAN MANHEIM
Emory University The purpose of the present study was to examine the impact of length of time of interaction and the individual’s type of interpersonal style on interactants’ attraction for one another. Thirty-nine dyads made up of female subjects interacted for either a brief time (5 min) or a relatively longer time (7.5 min). Based on a circumplex model of personality each subject was paired with a partner whose personality style was either complementary to or anticomplementary to her own. It was predicted and found that complementary dyads engaged in a higher number of verbal exchanges and preferred less interpersonal distance than anticomplementary ones. However, using a self-report measure, complementary relationships did not, as was predicted, lead to greater attraction than anticomplementary ones. The viability of reactive and nonreactive measures of relationship impact was discussed within the framework of interpersonal theory. o 19% Academic PXSS. IN.
Sullivan (1953) was among the earliest theorists to emphasize the importance of interpersonal factors in the development of personality. He believed that individuals used interpersonal means to achieve “satisfaction” of their basic biological needs and “security,” from the debilitating effects of anxiety. Individuals learned to foster the growth of interactions that led to satisfaction and security. Other theorists have attempted to define the kinds of interpersonal transactions associated with satisfaction and security, and their efforts have resulted in contemporary interpersonal theory (Andrews, 1989; Kiesler, 1988; Wiggins, 1982). Lear-y (1957) and Carson (1969) were instrumental in developing an initial circumplex model of interpersonal interactions that assumes interpersonal behavior is a blend of two primary dimensions: control (sometimes called dominance or status) and afjiliation (sometimes called friendliness or love). According to Kiesler (1988) “people interacting with each other continually are negotiating two major relationship issues: how friendly or hostile they will be with each other, Reprint requests should be addressed to Stephen Nowicki, Jr., Department of Psychology, Emory University, Atlanta, GA 30322. 322 W92-6566l91 $3.00 Copyright 8 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
INTERPERSONAL
COMPLEMENTARITY
323
and how much in charge or in control each will be in their encounters” (Pm 48). The circumplex model has been used to define complementary relationships or those interpersonal interactions that provide satisfaction and security for both participants. Complementary transactions take place when both parties confirm their self-presentations of control and affiliation. In complementary transactions there is reciprocity on the control dimension (i.e., dominance pulls for submission; submission pulls for dominance) and similarity on the affiliation dimension (friendliness pulls for friendliness; hostility pulls for hostility). Therefore complementary transactions occur with friendly dominant-friendly submissive and hostile dominant-hostile submissive pairings. In contrast, anticomplementary transactions are those in which self-presentations are completely negated on both dimensions and noncomplementary transactions are those in which only one of the two dimensions is affirmed. Empirical support for complementary transactions which seemed substantial at first (Kiesler, 1983) has been called into question. Orford (1986) after reviewing evidence for the complementary hypothesis, concluded that there was support for friendly complements (friendly dominant and friendly submissive) but not for hostile ones (hostile dominant and hostile submissive). He suggested caution in applying the complementary hypothesis uncritically and urged researchers to make explicit the impact of factors such as the length of time in the relationship. Paddock and Nowicki (1986) have criticized the usefulness of the most frequently employed method of assessing interpersonal styles, the selfreport approach (e.g., Interpersonal Check List). They suggested researchers use more comprehensive and direct means of assessing interpersonal style. However, even those researchers who have recently attempted more comprehensive ways of measuring interpersonal style (e.g.. Ickes, 1983) have failed to take into account the fact that interpersonal styles include both verbal and nonverbal components (Kiesler, 1991). In fact, as well as being an integral part of interpersonal styles, nonverbal behavior also may be a significant way the impact of an interpersonal interaction can be measured. That is, there may be a difference between what interactants are required to report feeling about their partners and how they actually feel about them. Since there is some reason to believe that nonverbal aspects of behavior may more accurately reflect how a person is actually feeling, they ought to be included as a measure of interpersonal impact (DePaulo et al. 1977). Regardless of how it is measured, the impact of complementary interpersonal styles, or the lack of them, needs to be viewed from the perspective of relationship development (e.g., Patterson, 1982). Duke and Nowicki (1982) proposed that relationships progressed through four
324
NOWICKI
AND MANHEIM
phases: choice, beginning, deepening, and termination. From their perspective, the negotiation of relationship definition is not as essential during the earliest stages of a relationship, but becomes more important as the interaction continues over time. Thus, complementary transactions should be less important and overlearned codes of manners more important when the interaction is a brief one between strangers. However, as a relationship continues past its early stage, Duke and Nowicki suggest that the importance of overlearned conventions fades and the basic negotiation of the relationship in terms of complementary transactions begins to take place. If Duke and Nowicki are correct, then complementary transactions ought to determine relationship valence only when interactions progress past the initial stage. In summary then, in the present study the authors sought to evaluate the assumption that complementary interpersonal styles should produce positive, security-enhancing transactions. Using a circumplex model and applying Duke and Nowicki’s assumptions concerning the development of relationships, it was predicted that the positive effects of complementary interactions would be reflected in a longer as opposed to a shorter term relationship. More specific to the present study, in a longer term interaction, subjects who were in complementary as opposed to anticomplementary relationships would show a higher number of verbalizations, prefer less interpersonal distance between themselves, and report more attraction toward one another. Further, to assess whether actual, and not just expected, interaction over time was needed for complementary interpersonal styles to have their impact on relationships, a condition was added in which subjects were led to believe that they were going to interact for a longer period of time, but in actuality only interacted for a brief time. METHOD Subjects Seventy-two female subjects were chosen from a pool of undergraduate students at Emory University. Subjects participated in this study as a requirement of their introductory psychology course. To be chosen to participate in the present study, subjects had to be congruent, that is, have similar verbal and nonverbal personality styles as indicated by being on the same side of the circumplex. To obtain the final 72 subjects, 104 subjects were assessed. While normative data are not available for the incidence of verbal and nonverbal congruence in interpersonal style, it was assumed that there was an equal distribution of congruence and incongruence in the population. About a fourth of the sample was classified as having incongruent verbal and nonverbal interpersonal styles. Since there are theoretical reasons for suggesting that incongruity may be associated with maladjustment (Kiesler, 1983), and the study of maladjustment was not the primary focus of the present study, it was decided to focus on just those subjects who presented congruent interpersonal styles. The final sample consisted of 22 friendly-dominant, 20 friendly-submissive, 16 hostiledominant, and 14 hostile-submissive subjects.
INTERPERSONAL
COMPLEMENTARITY
325
Procedure Subjects were initially contacted by phone or letter. They were told that all information would be kept confidential. Their next contact involved an interview that was videotaped. The video portions and verbal transcripts of the interview were rated by judges for verbal and nonverbal interpersonal style. Subjects continued in the study if they were rated as being congruent across verbal and visual channels with respect to status and atliliation. Once selected, subjects were assigned to complementary or anticomplementary dyads. For instance, if a subject’s interpersonal style was judged to be friendly-dominant and they were placed in the complementary condition they then were paired with someone who had a friendly-submissive interpersonal style; if she was placed in the anticomplementary condition she was paired with someone who was hostile-dominant. Each dyad was then assigned to one of the three interaction time conditions. Group 1 consisted of complementary and group 2 of anticomplementary dyads who were told that they would be interacting with their partner for 5 min and who actually did interact for 5 min. Group 3 and group 4 consisted of complementary and anticomplementary dyads, respectively, who were told they would interact with their partner for 75 min and who acutally did interact for that period of time. Finally, group 5 and group 6 consisted of complementary and anticomplementary dyads, respectively, who were told that they would interact with their partner for 75 min, but who actually interacted for 5 min. All groups contained 12 subjects each (6 dyads).
Selection Interview for Development of Selection Measure The verbal and nonverbal ratings of dominance-submission and friendliness-hostility were based on responses to a series of hypothetical situations designed to elicit interpersonal attitudes from respondents (e.g., Russell & Mehrabian, 1977). Five emotion-eliciting situations were used as the stimuli in the present study. Initially 10 situations were constructed and presented to students (n = 15) who were not participants in the present study. All situations were designed to reflect what was likely to occur in a female undergraduate’s life at college and included an interpersonal dilemma. For instance, students were asked in one situation to describe what they would do if they were asked out on a date after they had made plans with another friend. Written transcripts of responses were scored by trained raters. The five situations which elicited the strongest status and affiliation responses were selected for use in this study.
Application of Selection Measure to AssessInterpersonal Styles Interpersonal style rating scales for verbal and nonverbal channels. Rather than assuming that the subjects’ verbal self-report accurately reflected both the verbal and nonverbal aspects of their actual interpersonal style (as is the case with the self-report RX), an assessment strategy was employed that measured both verbal and nonverbal aspects of the communicative process. In this approach (Duke, 1978), a rater is asked to indicate the intensity of the status and the affihation of the subjects’ observed behavior on two 16O-mm lines anchored by descriptions of the two concepts. Duke (1978) reported significant, but not identical, correlations with ICL scores (r = .78 for status and r = .!?2 for affiliation). For the present study, two judges rated status and affiliation of verbal messages using a written transcript of taped interviews while two other judges rated the status and affiliation of nonverbal components using the video portion of the taped interviews. Prior to the actual scoring, the status and affiliation dimensions were explained to the judges using anchor scenes, scripts, and general descriptions of each interpersonal dimension. The explanations for the status and affiliation dimensions were based on the adjectives and descriptive phrases that anchor the control and affiliation dimensions that underlie circumplex models as well as the verbal and nonverbal cues which have been shown to
326
NOWICKI
AND MANHEIM
influence interpersonal behavior. Specifically, judges were instructed to use the adjective as guidelines for their ratings of the verbal content of the target person’s statements. If thl target person responded verbally to a particular hypothetical situation with an attitude o authority and leadership, then that person received a high rating of dominance. On the other hand, statements which reflected an unwillingness to take charge in certain situation: were classified as submission (Argyle et al., 1970; Heller, Myers, & Kline, 1963; Leary 1957). On the affiliation dimension, statements were classified as friendly if they conveyec a warmth and openness toward others. However, statements reflecting a lack of warmth 01 an expression of indifference toward others were rated as hostile (Bugental, Kaswan, & Love, 1970; Leary, 1957; Teyber, Messe, & Stollack, 1977). Two decision rules were included for the verbal response ratings during the training sessions. First, since the emotion-eliciting situations tended to evoke socially appropriate responses, the relative absence of friendly comments in a response resulted in ratings of hostility. Second, in terms of the status dimension, indications of equivocation (e.g., “may” or “might”) resulted in a rating of submissiveness. The guidelines for the visual channel ratings were drawn from the available research in the area of nonverbal communication. A dominant style was operationally defined by the following nonverbal cues: head slightly raised, direct body orientation, and arms and legs in a relaxed position. Submissiveness was defined by lowered head, nervous smile, and body tension. These criteria were based on investigations by Argyle et al. (1970), Goffman (1961) and Mehrabian (196Q Friendly nonverbal behavior was defined by frequent eye contact, forward trunk lean, smiling, and openness of arms. In contrast, hostile nonverbal behavior included infrequent eye contact, backward trunk lean, arms crossed, and unsmiling or strained smiling (Ekman & Friesen, 1975; Friedman, 1979; Gottman, Markman, & Notarius, 1977; Graves & Robinson, 1976; Mehrabian, 1972). A final verbal score was reached when both judges gave the target person the same status and affiliation ratings (on the same side of the midpoint and within 10 mm) on a majority of the responses provided by this person to the emotioneliciting situations. For example, if both judges rated four of five responses as dominant on the status dimension and four of five responses as friendly on the affiliation dimension, the target person received a friendly-dominant verbal score. A final nonverbal score was reached when both judges gave similar status and affiliation ratings to the video portion of the stimulus tape. The judges based their ratings on a S-min segment taken midway through the interview. The first 3 min of the interview were not included in the scoring procedure since the target person may have experienced some initial anxiety in the interview situation. In the event that there was disagreement between judges regarding one of the interpersonal dimensions, a third judge scored the necessary portion of the interview (transcript or video). If judges disagreed on both dimensions of the verbal or nonverbal channel, the target person was eliminated from the study (n = 2). For the nonverbal channels, judges rated the videotapes with the following percentages of agreement: Affiliation (friendly-hostile), 80%, and Status, (dominant-submissive) 74%. For the verbal channels, judges rated the transcripts with the following percentages of agreement: Affiliation, 74%, and Status, 80%.
Assignment
to Interacting
Dyaa!s
Cooperative tusks. Tasks were selected that invited partners to interact in a compatible manner (Ekstrand, 1980). Subjects interacting with one another were asked to (1) generate a list of words using the letters provided by a key word, (2) propose improvements and changes in campus life, and (3) devise a solution to an imaginary situation. Subjects in the expect 5 min of interaction/interact for 5 min condition (groups 1 and 2) were given the following instructions:
INTERPERSONAL
COMPLEMENTARITY
327
As I told you earlier, I am studying problem-solving strategies in women. Today, I will be asking you to work with your partner for a total of five minutes on a word task. I will give you a word and I want the two of you to use the letters in this word to make up new words. This is not a contest. I just want you to work together and come up with as many words as you can.
Subjects in the expect 75min interaction/interact for 75 min condition (groups 3 and 4) were introduced to each other and were given the same instructions except that they were told they would spend 75 min interacting with one another. After the dyads spent 5 min on the word task, the examiner returned and explained that the next task would involve deciding how to spend $1 million. Following the completion of this task, subjects were told to take a break before moving on to the next task. During this break, the dyads were observed through a one-way mirror and their verbal exchanges recorded. After the break, the experimenter returned to the room and read the instructions for the next task which was similar to the first task completed by the subjects. Following completion of this third task interaction the instructions for a fourth and final task were read. This time subjects were told they were going on a camping trip and needed to decide what to take along. Subjects in the expect a 75min interaction/interact for 5 min condition (groups 5 and 6) were given the same instructions as subjects in groups 3 and 4. However, after completing the 5-min word task, they were told that they had been placed in the wrong condition and that they actually were finished with the experiment. They then completed the ending procedures.
Dependent Measures Number of verbal exchanges. To indicate whether interactants tended to engage in or withdraw from an interaction, their verbal exchanges were counted during the unstructured 20-min break in the 75-min condition. A rater observed the dyads through a one-way mirror and gave one point to the dyad each time an interactant spoke and the other responded during a 30-s interval. The scores could range from 0 to 40. Interpersonal distance. A behavioral rather than self-report measure of interpersonal distance was selected for the present investigation because of the numerous limitations of the latter (e.g., Duke & Nowicki, 1972). Following the interaction and completion of the Interpersonal Judgment Scale (IJS) (see below) partners were asked to take a chair that was placed against the wall and move it next to their partner to discuss the experiment. When subjects left the room, the distance between the centers of the chairs was measured with a tape measure. Self-reported attraction. The IJS is a measure of initial attraction consisting of six items rated on a ‘I-point scale: intelligence, knowledge of current events, adjustment, and three attraction items (probable liking for the other, probable enjoyment of working with other, and probable enjoyment of having the other as a roommate) (Ettinger, Nowicki, & Nelson, 1970). The three attraction items are summed to yield a measure of attraction. Scores can range from 3 to 21. The higher the score, the greater the attraction. In all conditions while subjects completed the required questionnaires, chairs were arranged in another room so that they were equally spaced against the walls. The precise location for each chair was marked off by small pencil marks that were not noticeable to the subjects. At the proper time, subjects were told that a mixup in room assignments necessitated that they move from the experimental room to another one (where the chairs were located) to complete their forms.
328
NOWICKI
AND MANHEIM
TABLE 1 MEAN AND STANDARDDEVIATIONS OF INTERPERSONAL DISTANCE FOR COMPLEMENTARYAND ANTICOMPLEMENTARYDYADS WHO EXPERIENCEDONE OF THREE TIMED INTERACTIONS Time interactions Expect 5 min interact 5 min
Expect 75 min interact 5 min
Expect 75 min interact 75 min
Type of dyad
Mean
SD
Mean
SD
Mean
SD
Complementary Anticomplementary
56.17 54.50
11.37 15.35
55.33 53.67
12.97 12.47
54.00 70.00
15.41 5.52
RESULTS The means for the dependent measures are given in Tables 1 and 2 for the type of relationship (complementary versus anticomplementary) by interaction condition (expect a 5-minute interaction/interact for 5 minutes; expect a 75min interaction/interact for 75 min; expect a 75min interaction/interact for 5 min). Consistent with the behavioral procedures used to assess interpersonal styles, two of the dependent measures were behavioral and nonreactive: the number of verbal exchanges made during the break of the 75-min condition and the interpersonal distance the interactants created between themselves by placement of the chairs when they completed forms. The third dependent measure was a reactive selfreport questionnaire measuring self-reported attraction, the Interpersonal Judgment Scale. Data were subjected to 2 x 3 (complementary vs anticomplementary x three different time conditions) analyses of variance except for the measure of verbal exchanges which was only taken between groups 3 and 4. Collection of data for the first dependent measure, number of verbal TABLE 2 MEAN AND STANDARDDEVIATIONS OF INTERPERSONALJUDGEMENTSCALE SCORESFOR GXHPLEMENTARY vs ANTICOMPLEMENTARYDYAD~ IN THREE TIMED INTERACTIONS Time interactions Expect 5 min interact 5 min
Expect 75 min interact 5 min
Expect 75 min interact 75 min
Type of dyad
Mean
SD
Mean
SD
Mean
SD
Complementary Anticomplementary
15.00 16.92
3.30 1.93
17.00 16.67
2.00 1.07
14.42 15.08
3.23 4.12
INTERPERSONAL
COMPLEMENTARITY
329
exchanges, only took place during the break time in groups 3 and 4 where dyads interacted for 75 min. In terms of number of verbal exchanges, it was predicted and found that complementary dyads (mean = 29.17) produced significantly more verbal exchanges than did anticomplementary ones (mean = 7.00), t (11) = 3.38, p < .05. Interpersonal distance responses were obtained from every dyad and subjected to a 2 x 3 analysis of variance (see Table 1 for means). The complementary hypothesis was supported by those subjects who expected to interact for 75 min and who actually did interact for that period of time (groups 3 and 4). The complementary-anticomplementary dyads by type of interaction analysis of variance indicated a significant interaction between these two factors, F(1,33) = 4.17, p < .05. Post hoc comparisons revealed that in the 75-min condition complementary dyads (mean = 54 in.) placed less distance between themselves than did anticomplementary pairs (mean = 70 in.). No other significant differences in interpersonal distance were found between complementary and anticomplementary dyads in either of the 5-min interaction conditions. Although support was found for the positive impact of complementarity in the dyads that interacted for 75 min when nonreactive behavioral measures were used, that was not the case when self-reported attraction was employed. Complementary pairs (each subject of the pair contributed an attraction score) did not report higher interpersonal attraction for one another than did anticomplementary dyads (see Table 2 for means). There was, however, an unpredicted main effect for length of time dyads interacted, F(1, 69) = 3.72, p < .05). Post hoc testing for the source of the main effect for length of time dyads interacted indicated that subjects in the expect 75 min of interaction/interact for 5 min condition (groups 5 and 6) reported greater attraction for one another (mean = 16.83) than subjects in the expect 75 min of interaction/interact for 75 min (mean = 14.75). No other comparisons were significant. Not surprisingly, in light of the pattern of results, the dependent measures were inconsistently related to one another. The two nonreactive, behavioral measures (number of verbal exchanges and interpersonal distance) were significantly related to one another, r(12) = - 44, p < .05, but self-reported attraction scores were unrelated either to interpersonal distancing, r(72) = .Ol, p > .05, or to number of verbal exchanges, r(12) = .ll, p > .05. DISCUSSION In the present study, the authors used directly observed interpersonal interactions to evaluate a circumplex model operationalization of Sullivan’s complementary hypothesis. It was proposed that predictions gen-
330
NOWICKI
AND MANHEIM
erated for the complementary hypothesis would be most clearly shown within the context of an ongoing and more lengthy interaction. Results generally supported predictions when nonreactive, behavioral means of evaluation were used to assess longer as compared to shorter interactions. The positive impact of complementary interpersonal styles was reflected in their greater number of verbal exchanges and a closer interpersonal distance when interacting as compared to anticomplementary pairings. These results are consistent with the view that complementary personality styles elicit behavior from others that increase feelings of security. However, because the self-report measure of attraction and the behavioral measure of interpersonal distance, but not the number of verbal exchanges, were gathered late in the interpersonal process, it is difficult to determine when complementary transactions may begin to impact on an ongoing relationship. In fact, investigators have yet to determine how much time is required for complementarity to make its impression on a relationship. Kiesler (1983) stated that “it is unclear how interpersonal complementarity applies over the temporal range of continuing transactions between interactants” (p. 209). It is possible that the lack of consistent empirical support for the complementary hypothesis reported by Orford (1986) may be due, in part, to an inability to pinpoint precisely when interactants are engaged in transactions which are affected by their complementary personality styles. Consistent with the results of the present study, investigators who have found support for the complementary hypothesis are usually those who went beyond a single brief interaction or studied interpersonal behavior over differing lengths of time (e.g., Dietzel & Abeles, 1975; Rausch, 1965; Shannon & Guemey, 1973). There was support for the positive impact of complementary interpersonal styles when it was measured by the number of verbal exchanges and interpersonal distance, but not when self-reported attraction scores were used. Perhaps self-report scores were not higher for complementary as opposed to anticomplementary dyads because they are more likely to be affected by rules of social interaction than behavioral measures. Even after 75 min of interaction, it may be difficult for individuals to report publicly that they do not like another person. Further, self-report scores may be susceptible to the effects of social desirability. Regardless of how one might actually feel, there may be a tendency to report liking a stranger with whom you have just interacted for a time. Another unpredicted finding provided further support that self-report measures may not accurately reflect attraction related to the presence of complementary interpersonal styles. It was predicted that subjects who were led to expect a 75-min interaction would respond differently to one another than those subjects who expected to be engaged and were engaged in a 5-min interaction. This hypothesis was based on the possibility that
INTERPERSONAL
COMPLEMENTARITY
331
complementarity would lead to a positive interpersonal interaction only within the context of an actual and prolonged interaction. If partners did not have the opportunity to accumulate interpersonal experiences, it was expected that their behavior would be determined by manners or overlearned rules of interaction, rather than by the presence or the absence of complementary interpersonal styles. However, the main effect found for interaction time on self-reported attraction indicated that subjects who expected to, but did not engage in, a 75min interaction reported their partners to be more attractive to them than subjects who expected to and did interact for 75 min. An important difference between these two groups is that one group completed the attraction scale with the expectation of a future interaction with a partner, while the other group completed the scale following the end of an interaction. This is consistent with results of studies reviewed by Byrne and Griffitt (1973) in which it was found that attraction was greater toward subjects who were ambiguously or even negatively described when future contact was expected than when such contact was not expected. It could be that the attraction scores reflected subjects’ anticipated attitudes or use of social convention, rather than their true reaction to the emotional engagements imposed on them by their partners. In addition, self-reported attraction may have functioned as an expectancy of where interactants hoped the relationship was heading rather than where it actually was. The lack of relation between nonreactive behavioral and self-report measures of attraction also deserves some comment. Both are sometimes used interchangeably by those who study relationships in spite of the fact that researchers have cautioned that they may be measuring different facets of attraction (Byrne & Griffitt, 1973). It may be that these measures reflect different aspects of the interpersonal process. As well as attempting to clarify what measures should be used to evaluate relationships, researchers need to generalize their work to all kinds of dyads. In the present study, only females were used and their results may not be generalizable to males or to mixed-sex dyads, especially those with verbalizations which were only evaluated for 12 dyads. Likewise, only complementary and anticomplementary relationships were examined. Carson (1%9) also described noncomplementary relationships in which one of the two major dimensions was complementary. He believed that this was probably the most frequent type of interaction and because of that was deserving of study. REFERENCES Andrews, J. D. (1989). Integrating visions of reality: Interpersonal diagnosis and the existential vision. American Psychologist, 44, 8034317. Argyle, M., Salter, V., Nicholson, H., Williams, M., & Burgess, P. (1970). The commu-
332
NOWICKI
AND MANHEIM
nication of inferior and superior attitudes by verbal and nonverbal signals. British Journal of Sociology and Clinical Psychology, 9, 222-231. Bugental, D. E., Kaswan, J. W., & Love, L. R. (1970). Perception of contradictory meanings conveyed by verbal and nonverbal channels. Journal of Personality and Social Psychology, 16, 647-655. Byrne, D., & Griffitt, W. (1973). Interpersonal attraction. In P. H. Musser & M. R. Rosenxweig (Eds.), Annual review. Palo Alto, CA: Annual Reviews. Carson, R. C. (1969). Znteraction concepts of personality. Chicago: Aldine Pub. DePaulo, B. M., Rosenthal, R., Eisenstat, R. A., Rogers, P. L., & Finkeistein, S. (1978). Decoding discrepant nonverbal cues. Journal of Personality and Social Psychology, 36, 313-323. Dietzel, C. S., & Abeies, N. (1975). Client-therapist complementarity and therapeutic outcome. Journal of Counseling Psychology, 22, 264-272. Deutsch, M. (1949). An experimental study of the effects of cooperation and competition upon group process. Human Relations, 2, 199-231. Duke, M. P. (1978). A new method for measuring the circumplex. Paper presented at the Southeastern Psychological Association Meetings, Atlanta, GA. Duke, M. P., & Nowicki, S. (1972). A new measure and social-learning model for interpersonal distance. Journal of Experimental Research in Personality, 6, 119-132. Duke, M. P., & Nowicki, S. (1982). A social learning theory conceptualization of interpersonal relationships. In D. Kiesler & J. Anchin (Eds.), Handbook of interpersonal relationship. New York: Pergamon. Ekman, P., & Friesen, W. (1975). Unmasking the face. Englewood Cliffs, NJ: PrenticeHall. Ekstrand, M. (1980). Differential effects on interpersonal comfort of competition and cooperation among college students with varying interpersonal styles. Unpublished Honors thesis, Emory University. Ettinger, R., Nowicki, S., & Nelson, D. (1970). Interpersonal attraction and the approval motive. Journal of Experimental Research in Personality, 4, 95-99. Friedman, H. (1979). The interactive effects of facial expressions of emotion and verbal messages on perceptions of affective meaning. Journal of Experimental Social Psychology, 15, 453-449. Goffman, E. (1961). Encounters. Indianapolis, IN: Bobbs-Merrill. Gottman, J., Markman, H., & Natarius, C. (1977). The topography of marital conflict: A sequential analysis of verbal and nonverbal behavior. Journal of Marriage and the Family, 461-477. Graves, J. R., & Robinson, J. D. (1976). Proxemic behavior as a function of inconsistent verbal and nonverbal messages. Journal of Consulting Psychology, 23, 333-338. Heller, K., Myers, R. A., & Kline, L. V. (1963). Interviewer behavior as a function of standardized client roles. Journal of Counseling Psychology, 27, 117-122. Ickes, W. (1983). A basic paradigm for the study of personality, roles and social behavior. In W. Ickes & E. S. Knowles (Eds.), Personali@, roles, and social behavior (pp. 305341). New York: Springer-Verlag. Kiesler, D. J. (1983). The 1982 Interpersonal Circle: A taxonomy for complementarity in human transactions. Psychological Review, 90, 185-214. Kiesler, D. (1986). Interpersonal methods of diagnosis and treatment. In J. 0. Cavenar (Ed.), Psychiatry (Vol. 1, No. 4, pp. l-23). New York: Lippincott. Kiesler, D. (1988). Therapeutic metacommunication: Therapist impact disclosure as feedback in psychotherapy. Palo Alto, CA: Consulting Psychologist Press. Kiesler, D. (1991). Interpersonal methods of assessment and diagnosis. In C. R. Snyder &
INTERPERSONAL
COMPLEMENTARITY
333
D. R. Rorsyth (Eds.), Handbook of social and clinical psychology: The health perspective. Elmsford, NY: Pergamon. Leary, T. (1957). Interpersoruddiagnosi of personality: A functional theory and methodology for personality evaluation. New York: Ronald. Mehrabian, A. (1968). The inference of attitudes from the posture, orientation, and distance of a communicator. Journal of Consulting Psychology, 32, 296-308. Mehrabian, A. (1972). Nonverbal communication. Chicago: Aldine. Noller, P. (1985). Nonverbal communication and ma&al interaction. New York: Pergamon. Oxford, J. (1986). The rules of interpersonal complementarity: Does hostility beget hostility and dominance, submission? Psychological Review, 93, 365-377. Paddock, J. R., & Nowicki, S., Jr. (1986). An examination of the Leary circumplex through the Interpersonal Check List. Journal of Research in Personality, 20, 107-144. Patterson, M. L. (1983). A sequential functional model of nonverbal exchange. Psychological Review, 89, 231-249. Rausch, H. L. (1%5). Interaction sequences. Journal of Personality and Social Psychology, 2, 487499. Russell, J. A., & Mehrabian, A. (1977). Evidence for a three-factor theory of emotions. Journal of Research in Personality, 11, 273-294. Shannon, J., & Guemey, B. (1973). Interpersonal effects of interpersonal behavior. Journal of Personality and Social Psychology, 101,53-68. Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York: Norton. Teyber, EC., Messe, L. A., & Stollack, G. (1977). Adult responses to child communications. Child Development, 48, 1577-1582. Wiggins, J. S. (1982). Circumplex models of interpersonal behavior in clinical psychology. In P. C. Kendall & J. N. Butcher (Eds.), Handbook of research methods in clinical psychology. New York: Wiley.